Agricultural Water Management 131 (2014) 70–78
Contents lists available at ScienceDirect
Agricultural Water Management
j ournal homepage: www.elsevier.com/locate/agwat
Does output market development affect irrigation water institutions?
Insights from a case study in northern China
a,b,∗ c a,b a
Lei Zhang , Xueqin Zhu , Nico Heerink , Xiaoping Shi
a
China Centre for Land Policy Research, Nanjing Agricultural University, Nanjing, PR China
b
Development Economics Group, Wageningen University, Wageningen, The Netherlands
c
Environmental Economics and Natural Resources Group, Wageningen University, Wageningen, The Netherlands
a r t i c l e i n f o a b s t r a c t
Article history: The main aim of this paper is to examine the impact of changing external conditions on irrigation water
Received 16 November 2012
institutions in northern China. To this end, we perform a case study analysis of the impact of output
Accepted 16 September 2013
market development on irrigation water transactions, using survey data collected among 315 households
Available online 21 October 2013
in Minle County, Zhangye City, Gansu Province, covering the year 2009. Households in this region possess
tradable water use rights. Moreover, a major agro-processing company has recently been established and
Keywords:
the local government intervenes in the allocation of water to stimulate farmers to grow a cash crop for
Irrigation water
that company. Despite these favourable enabling and driving factors, we find that market water trade is
Institutional change
virtually absent. Instead, we observe that reciprocal water use arrangements (water swaps) have emerged
Output market
at a limited scale. We argue that factors other than an improvement in the output market (such as producer
Transaction costs
China ignorance, centrally set prices, trust) need to be considered, if improvement in the market for irrigation
water is to occur.
© 2013 Elsevier B.V. All rights reserved.
1. Introduction distribution and an innovative system of water resources property
rights allocation and trading.
China is a country with substantial water resources, but their Zhang (2007) and Zhang et al. (2009) examine the water prop-
regional distribution is highly unequal. Water availability in the erty rights system that was implemented in Zhangye City. These
3
north (757 m per person in 2003) is almost 25% below the inter- studies find that high transaction costs in some parts of the region,
3
nationally accepted water scarcity threshold of 1000 m per person, and management, legal, administrative and fiscal barriers in cases
3
while water availability in the south (3208 m per person in 2003) where transaction costs are low, discourage farmers from saving
is relatively abundant (Shalizi, 2006). and trading surplus water. As a result, trading of water use rights
The water resources available for agricultural production in is almost non-existent in this pilot project area.
China are rapidly declining due to increased water demand for Induced institutional innovation theory suggests that new
industrial use and household consumption. The use of water in institutions, such as tradable water use rights and non-market
agriculture as a share of total water use has steadily declined from institutions, may emerge when resources become more scarce due
around 80% in 1980 to 61.3% in 2011 (Shalizi, 2006; National Bureau to growing population density, commercialization of agriculture,
of Statistics of China, 2012). Technical innovations as well as water or exogenous technological change (Hayami and Ruttan, 1985;
policy and management reforms are required to improve water use Platteau, 1996). Although the theoretical literature elaborating the
efficiency in agriculture to meet growing food demands (Rosegrant gains from institutional changes is vast and growing (Bromley,
and Cai, 2002; Yang et al., 2003). The Ministry of Water Resources of 1989; Saleth and Dinar, 2000), empirical studies examining drivers
the PR China has initiated a number of pilot projects to gain expe- of institutional change are scarce due to lack of suitable data sets.
rience with the development of water-saving irrigation systems. Appropriately chosen case studies can provide deeper insights into
The first of these pilot projects was initiated early 2002 in Zhangye the role of changing external conditions in stimulating institutional
City, an oasis with rich agricultural resources in Gansu Province change, and may be used to formulate hypotheses on driving forces
in northern China. Measures taken under this project include the of institutional change that can be tested at a larger scale.
construction of an engineering system that optimizes the water In Minle County, one of the six counties in Zhangye City, a large
potato processing factory was established in 2008. The factory is
owned by Aviko Gansu Potato Processing Co., Ltd., a joint venture
∗ of Aviko – one of the four largest potato processing companies in
Corresponding author. Tel.: +31 647754258; fax: +31 20 4852623.
E-mail address: zhanglei [email protected] (L. Zhang). the world – and the local government of Minle County. To meet
0378-3774/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.agwat.2013.09.008
L. Zhang et al. / Agricultural Water Management 131 (2014) 70–78 71
the demand of this factory, the area grown with potatoes in Minle as well-defined water rights and water markets) are required to
County is rapidly being expanded at the instigation of the local achieve an efficient allocation of water over its users such that the
government. Potatoes need a relatively large amount of water, total net benefits of water are maximized. Water institutions can
but the water should be applied at a later stage in the season than be defined as the humanly devised constraints that regulate water
many other crops grown in the region. A detailed examination development, allocation and utilization. Different institutions are
of the changes in the allocation of water to farm households and combined in reality for water management, and continued public
the trading of water by households that occurred since 2008 in sector participation is required to deal with the common property
1
Minle County may add to a better understanding of the impact of character of water and to address externalities (Griffin, 2006). As
output market development on water institutions. Given the fact a result, various types of water institutions have been established
that Minle County is located within the water-saving pilot area of in different areas around the globe.
Zhangye City, such research may also provide important insights According to the first welfare theorem, Pareto efficient allo-
into further policy reforms that are needed for establishing an cations of water can be achieved by establishing water property
efficient system of water resources property rights allocation and (use) rights and water markets, provided transaction costs are zero
trading. and a number of additional conditions, such as absence of exter-
The objective of this paper is to examine the changes in water nalities, are satisfied. A resource being managed as a transferable
institutions that took place in Minle County, northern China property will cause a market to arise and the market will produce a
after the establishment of a large potato processing company resource-conserving signal, namely its price (Griffin, 2006). When
in 2008, and the driving forces of these changes, and to use the individual agents possess property rights in (natural) water, they
resulting insights to formulate policy recommendations on ways will be able to exchange water for money or other property.
to improve the functioning of water institutions. To this end, we Water trading means the exchange of water rights by willing
use data collected for the year 2009 among 315 households to buyers and sellers. Water trading is a scarcity-addressing strat-
assess the frequency of water exchanges after the company was egy to achieve Pareto efficiency because water can be used to its
established, and to examine factors affecting water exchanges highest value, when the conditions under which the first welfare
between households. We find that despite the development of the theorem holds are met (e.g. Zhu and Van Ierland, 2012). Economic
output market, no significant water trading emerged. Informa- theory suggests that, in a perfect market with full information, trad-
tion asymmetry between government and water users severely ing of water takes place until the marginal net benefits of all users
constrains the water use rights exchanges in the region, while are equalized. When a water trading scheme is implemented, the
low levels of non-kinship trust among villagers entail that most amount of water being transferred therefore depends on the differ-
observed exchanges take the form of water swapping instead of ences between the marginal net benefits of different users. With a
market exchanges. We conclude that without addressing these relatively large difference in marginal net benefits, water users are
bottlenecks, output market development is unlikely to boost the expected to trade water (transfer water rights). If there exist only
development of a tradable water use rights markets. small differences between the marginal net benefits, the traded
In the next section we present the theoretical framework, amounts are expected to be small.
focusing in particular on efficiency gains obtained by market and The existence of imperfect information in water market opera-
non-market water institutions, the role of transaction costs, and tions, however, contributes to high costs of searching, bargaining
the impact of exogenous and endogenous factors on water man- and other transaction costs that can pose a serious hurdle for direct
agement institutions. Recent developments in irrigation water market exchanges. Under such conditions, water transactions may
management in China are briefly summarized in Section 3, while take place through non-market institutions. The transaction cost
the research area (Minle County, Zhangye City) and the data col- approach in the so-called new institutional economics (NIE) pro-
lection method are introduced in Section 4. In Section 5, we use vides an appropriate tool to understand market and non-market
the survey data and insights gained through informal field visits exchanges under non-zero transaction costs (Williamson, 1979,
to examine water exchanges that occurred in the year 2009 and to 2007; Jia and Huang, 2011). Several basic forms of (market and
explain the very limited development of market and non-market non-market) water transactions can be distinguished, including
water institutions in the region. The conclusions of our study and exchanges in kind, temporary rentals, permanent sales of rights,
recommendations for further research and for policy making in this and various forms of option contracts (Young, 1986).
field are presented in Section 6. From the transaction costs perspective, the difference between
two traders’ marginal benefits of water evaluated at their initial
holding levels must be large enough to offset the marginal trans-
2. Theoretical framework
action costs involved in water trading under market institutions.
Water is used for many purposes such as irrigation in agri- If transaction costs are high, especially when they exceed the dif-
culture, hydropower generation, domestic consumption, industrial ferences in marginal benefits for many potential traders, they may
use and for environmental purposes. Water has an economic value become an obstacle to water trading. In such cases, non-markets
in all its competing uses and should therefore be treated as an eco- institutions can serve as market substitutes for better resource
nomic good (ICWE, 1992). Due to its physical attributes, however, allocation by economizing on transaction costs; well-known exam-
natural water is not a standard (private) economic good. Due to its ples of non-market institutions that have developed to reduce
fluid nature, exclusion is frequently impossible or may be obtained transaction costs in insurance, credit and labour markets include
at high costs. The consumption of water is considered by humans as share cropping, contract farming, and reciprocal labour sharing
non-rival and non-exclusive when it is available in abundant quan- (Hubbard, 1997; Gilligan, 2004; Williamson, 2007; Jia and Huang,
tities. It stops being a pure public good when the consumption or 2011; Takasaki et al., 2012). In transaction cost economics, “econ-
use by one person affects the utility or production possibilities of omizing on transaction cost is taken to be the cutting edge, [. . ..]:
others. But, like many other environmental resources, it tends to
remain non-exclusive long after it first became rival (Ellis, 1993,
pp. 259–260).
1
In this study, we focus on the functioning of a water market which is an impor-
With rising water scarcity, due to population growth, economic
tant element of the water-saving pilot project in Zhangye City. Potential externalities
development or other factors, the need for social investments
such as salinity of water are neglected, because the main water source in this region
in barriers to access rises. Appropriate water institutions (such is surface water.
72 L. Zhang et al. / Agricultural Water Management 131 (2014) 70–78
Transactions, with different attributes, are aligned with governance on the functioning of water institutions. Water scarcity, arising e.g.
structures, which differ in their cost and competence, so as to effect from competing uses of water, creates an endogenous pressure for
a transaction cost economizing outcome (Williamson, 2007, p. 17). change, inducing change in the performance of water institutions
An important institution that receives much attention within and thus water-using sectors (Saleth and Dinar, 2004). In locations
NIE is that of contracts. Contracts can be defined as the means where market exchanges are novel or infrequent, transaction costs
whereby parities design transactions to their mutual advantage can be especially high due to a lack of familiarity either by mar-
(Hubbard, 1997). Contracts can be formal and informal. An example ket participants, their legal representatives, or the administrative
of (usually informal) contracts are agricultural labour exchanges in agency (Griffin, 2006).
which multiple farmers temporarily pool their labour into teams The performance of water institutions is also affected by exoge-
and complete a task on each team member’s land in succession. nous factors such as historical forces, political arrangements,
These strict reciprocal labour-sharing arrangements without pay- demographic conditions, resource endowments, and economic
ments allow working capital constrained farmers in developing development. Economic policies, especially macroeconomic and
countries to increase their productivity in the context of credit trade (export market) reforms, also play an important role in pro-
and labour market imperfections (Gilligan, 2004). Incentive and/or viding impetus for institutional changes within the water sector
enforcement problems may be mitigated in team work through (Saleth and Dinar, 2004).
reputation mechanisms to secure future labour on the own farm, Given that various (endogenous and exogenous) factors influ-
mutual monitoring, social norms and peer pressure. In a case ence the performance of water institutions, it follows that similar
study on forest clearing among shifting cultivators in the Peru- water institutions that operate in different environments may
vian Amazon, both the efficiency and productivity of labour-sharing greatly differ in their performances. It also means that the actual
institutions is found to be higher as compared to labour mar- performance of an existing water institution in a given setting is an
ket exchanges under credit market imperfections (Takasaki et al., empirical question. Gaining insights into the most important fac-
2012). tors explaining institutional change and performance in the water
Likewise, a theoretical justification exists for reciprocal water- sector is not only relevant from a scientific point of view, but may
sharing use arrangements because of its Pareto efficiency. also contribute to the design of policies that stimulate a more effi-
Analogous to reciprocal labour sharing arrangements, smallholder cient use of limited water resources. This study intends to add to the
farmers in thin agricultural water markets may use reciprocal water literature in this field by performing a case study of the impact of
use arrangements to realize efficiency gains in water use while agricultural output market development on the development and
economizing on the transaction costs involved in water trading. To performance of water market exchanges and reciprocal water use
our knowledge, this alternative to market exchanges of irrigation arrangements.
water has not been examined in the literature so far.
An important factor affecting transaction costs in market and
3. Irrigation water management in China
non-market exchanges is trust. In the absence of sufficient trust,
market participants must protect themselves against moral hazard
Before the agrarian reforms in China in the late 1970s, water
or even theft through monitoring, contract writing and other costly
resources were managed primarily through collective owner-
initiatives, and should be prepared to litigate in case of default. In
ship arrangements. Since the start of the reforms, a variety
low-trust societies, people are therefore more likely to use infor-
of institutional arrangements have been established to govern
mal arrangements, trading only with people they know and relying
water resources (Mukherji and Shah, 2005; Zhang et al., 2008).
on informal punishment mechanisms (Easterly, 2005). Supportive
Besides contracting out of water management and joint manage-
empirical evidence is provided by Tu and Bulte (2010) in a study
ment through water users associations (WUAs), recent changes
of rice farmers in southeast China. Making a distinction between
in irrigation water management in China mainly involve tradable
local trust (trust among friends and family) and general trust (trust
water use rights and introduction of water pricing (Qu et al., 2011;
towards the community more broadly defined—encompassing a
Wang et al., 2010).
greater set of potential trading partners), their study finds that high
The establishment of water markets was made possible by the
levels of general trust are associated with the use of labour markets,
revised national Water Law that came into force in late 2002
while high levels of local trust are associated with a greater proba-
(Yuan and Chen, 2005). However, water markets in China are
bility to engage in reciprocal labour sharing arrangements. By the
at an elementary stage and are generally occurring outside of a
same token, general and local trust can be expected to play impor-
structured trading framework (WET, 2007). Examples of water use
tant roles in household decisions on using either formal water
rights (WUR) exchange to date mainly include sales from one local
markets or informal non-market exchange mechanisms.
government to another, and the transfer of WUR from irrigation
Efficiency gains of different market and non-market water insti-
districts to industries following water efficiency initiatives (Speed,
tutions are achieved through changes in the production costs
2009). Notably, these have been driven by the relevant govern-
(transformation costs) of water users (Saleth and Dinar, 2004;
ments and not by the free market (Speed, 2009).
Griffin, 2006; Zhu and Van Ierland, 2012). The type of water insti-
Water was generally considered a free good until the start of the
tutions in place greatly affects the efficiency of water allocation.
market-oriented agricultural reforms at the end of the 1970s. Since
The performance of water institutions can therefore be viewed
then the central government encouraged the adoption of a system
from the perspective of their functioning, i.e. operations such as
of volumetric surface-water pricing because prices may provide
active transfers of water rights. Various factors affect this rela-
incentives for using water more efficiently (Lohmar et al., 2003; Qu
tionship between the performance of water institutions and the
et al., 2011). Water fees were gradually introduced and increased
efficiency gains of water users. Previous studies make a distinction
in an effort to meet the cost of water supply and improve water
between endogenous factors, such as water scarcity and financial
efficiency. Current prices charged for irrigation water, however, are
constraints, and exogenous factors, such as macroeconomic reform,
generally believed to be well below levels that are efficient (i.e. that
political reform, international agreements, natural calamities, and
markets would set). Irrigation water prices often do not even cover
technological progress (e.g. Saleth and Dinar, 2000).
the costs of operating and maintaining irrigation systems (Hussain,
With respect to endogenous factors, the relative scarcity of
2005; Wang et al., 2004; Yang et al., 2003). This may hinder the
water and the transaction costs required to enforce water rights
efficient allocation of water under the prevailing water institutions.
and establish water markets are found to have significant impacts
L. Zhang et al. / Agricultural Water Management 131 (2014) 70–78 73
4. Description of the research area and data collection sub-divided into water user groups (WUGs), consisting of house-
holds having plots along the same channel. Since the plots of
In our research we use information that we collected via a house- different households within a WUG are irrigated at the same
hold survey held in Minle County, Zhangye City, Gansu Province. In time, households belonging to a WUG need to coordinate their
this section we first introduce the research area and then briefly planting decisions and water demands. Water use is measured at
discuss the method of data collection. the entrance of the second-level canals, and corrected for esti-
mated losses that occur between the point of measurement and
4.1. Research area the destination point. The water price charged was approximately
3 3
0.09 RMB/m in 2009 . Households can trade their WUR with other
Zhangye City is an oasis located midstream of the Heihe River, an
households, provided the price that is charged does not exceed
inland river that flows across Qinghai Province, Gansu Province and 3
0.3 RMB/m .
Inner Mongolia Autonomous Region. The river originates from the
The new potato processing company that was established in
Qilianshan Mountains in Qinghai province, crosses the Hexi corri-
Minle County in 2008 demands a specific variety of potatoes, named
dor in Gansu Province, and ends in Juyanhai Lake in Inner Mongolia.
Atlantic potatoes, for processing into flakes and starch. In order to
In the midstream of the Heihe River watershed, the land is flat, sun-
meet the growing demand for Atlantic potatoes, the local govern-
shine is abundant, and annual precipitation is very low while the
ment (which owns a major stake in the new factory) assigned quota
evaporation is high. But due to the availability of irrigation water
for Atlantic planting areas to lower level governments. And it also
from the Heihe River, the area has become a major grain and veg-
ordered the WMBs (the ‘sellers’ of irrigation water) in the region to
etables production base in Gansu province.
give priority to land planted with Atlantic potatoes in the allocation
According to the Ministry of Water Resources (2004), Zhangye
of water to the WUAs with in their districts.
City is severely short of water resources, even though it uses up
The water allocation priority policy requires that in spite of the
almost all the water of the Heihe River. Only 50% of the farmland
water scarcity in the region, a sufficient amount of irrigation water
is well irrigated, and much arable land has been abandoned due to
has to be reserved for irrigating Atlantic potatoes. The remaining
water shortage. Agriculture accounts for approximately 95% of all
quantity of irrigation water is allocated to land planted with other
water use and almost all water in the Heihe River is extracted for
crops. Specifically, water is allocated to Atlantic potatoes according
irrigation use. As a result, too little water flows into Juyanhai Lake.
to its actual planting area, while water allocation to other crops is
2
The lake dried out in 1992 and an area of 200 km around the lake
based on the WUR area. Normally the WUR area is smaller than the
became a desert (Zhang et al., 2009). 4 5
actual growing area . Moreover, one to three rounds of irrigation
Minle County is located between the foothills of the Qilianshan
are carried out for most crops, while four or five rounds are carried
Mountains and the lower lying Hexi corridor. Its total cultivated
out for Atlantic potatoes.
2
land area equals 860,000 mu , with irrigated land constituting
67%. Major crops in Minle County include barley, wheat, maize,
4.2. Data collection
sesame, rapeseed, garlic and potato. Surface water is the major
water resource for irrigated agriculture in the area. According to
A household survey, covering the year 2009, was held in Minle
the Water Bureau of Minle County, the use of groundwater is less
County in May 2010. One of the purposes of the survey was to
than 5% of total water use in irrigated agriculture due to the high
examine the changes that took place in irrigation water allocation
costs of pumping water from the wells.
and water markets since the establishment of the new company
Precipitation is Minle County, as in other parts of Zhangye City,
and the concomitant expansion of the potato market. In a base-
is unevenly distributed. The wet season is from May to September;
line survey, covering the year 2007 and carried out in May 2008,
it provides around 80% of yearly rainfall. The dry season is from
irrigating households were found not to undertake any (market or
October to April. Irrigation in Minle County is carried out from
non-market) water transactions.
the beginning of May to late-August. The period until June has a
The household survey in May 2010 was carried out by staff and
peak demand for water, but the rainfall during those two months is
students from Gansu Academy of Social Sciences, Gansu Agricul-
limited. These two months are known by local farmers in Zhangye
tural University, and Nanjing Agricultural University. The dataset
City as the ‘choked period’ (Chen et al., 2005).
includes information about, among others, crop production, use
The water used for surface irrigation in Minle County is stored in
of water and other inputs, water transactions, WUA participation,
seven reservoirs in the Qilianshan mountains, serving five irrigation
water and other prices, land tenure and land use. To ensure that all
areas within Minle County. A county-level water management
townships would be equally represented, the population in Minle
bureau (WMB) supervises the water allocation institutions within
County was stratified into 10 townships. Ten percent of the villages
the region.
in each stratum were randomly selected, giving a total sample of
Five lower-level WMBs, one for each of the five irrigation areas,
21 villages. In each of the 21 selected villages, 15 households were
arrange the water allocation to WUAs within their own irrigation
randomly selected to be interviewed. This gave us a dataset con-
area. A standard water allocation quota is assigned to the so-called 6
taining 315 observations (see Wachong Castro et al., 2010, and Ma
WUR land of the WUAs within the irrigation area served by a WMB.
et al., 2013, for more information).
Not all the irrigated land is classified as WUR land. Its size depends
Table 1 presents background information for some key variables
on the labour that was provided by a village to the construction
that were investigated in the survey. The average land holding size
of the reservoirs, the WUR land obtained by a WUA through auc-
tions, and other factors. Based on the actual availability of water in
a reservoir in each year, a certain percentage of this standard water
3
1 USD = 6.83 RMB (2009).
quota is actually distributed to the WUAs within the irrigation area 4
The average proportion of WUR land in total arable land is 72.6% for all house-
of a WMB.
holds in our sample.
5
WUAs are responsible for arranging the water allocation to Depending on their altitudes, some villages receive one or two rounds of
its member households. The households within each WUA are irrigation, while others receive three rounds.
6
Out of the 315 households that were interviewed in the 2009 survey, 265 house-
holds were also interviewed in the baseline survey covering the year 2007. The other
50 households could not be found, and were replaced by other randomly selected
2
One mu is equal to 1/15 ha. households within the same village.
74 L. Zhang et al. / Agricultural Water Management 131 (2014) 70–78
Table 1
Descriptive statistics of some key variables in the household survey.
Agricultural production inputs
Variables Unit No. of observ. Mean Std. Dev. Min Max
Land mu 312 19.6 11.1 1.60 71.3
Labour days/mu 310 8.29 6.90 0.80 58.9
3
Water m /mu 308 483 220 53.3 1420
Fertilizer jin/mu 312 208 77.9 33.5 626
Seed jin/mu 309 74.8 42.1 7.67 381
Age of head years 315 46.4 10.2 23 78
Education of head years 314 7.52 3.51 0 15
Total production value RMB/mu 302 33.3 33.2 0.5 256.5
Note: 1 jin = 0.5 kg.
Table 2
of the surveyed households equals 1.3 ha, about twice the average
Water transactions in Hongshuihe irrigation area in 2003, 2007 and 2009.
size of landholdings in rural China (which equals 0.6 ha according
to the 2006 Agricultural Census). There exists considerable vari- 2003 2007 2009
ation in landholding sizes, ranging from 0.1 ha for the smallest
Number of households interviewed 380 105 105
holding to 4.8 ha for the largest one. Average irrigation water use Number of households with water transactions 5 0 7
3 3
equals 483 m /mu. It varies from 53.3 to 1420 m /mu in the sam- Percentage of households with water transactions 1.3 0 6.7
ple. Fertilizer use is high, like in many parts of rural China with
Sources: Zhang et al. (2009) for 2003; our surveys for 2007 and 2009
predominantly irrigated agriculture. The average fertilizer use for
the households in our sample amounts to 1560 kg/ha.
Table 1 According to the Minle County Statistical Yearbook overall the percentage of households involved in water transactions
2009, the amount of rainfall during May-August 2009 was equal to remains very small.
218.3 mm, with the largest amount of rainfall in August: 107.6 mm. Water transactions are closely related to land rental transactions
8
Precipitation during the survey year is considered as average by the among households, since rented land also needs water . When land
WMB of Minle County. is being rented out to other households, the water rights belonging
to that land are normally part of the same deal. In our analysis
we disregard these so-called long term trades, as defined by Zhang
5. Water allocation and water transactions in Minle County
et al. (2009), consisting of water and land use rights that are being
transferred together. For the purpose of our analysis it is water
Out of the 315 households that we interviewed in the survey,
transfers without parallel land transfers that matters.
15 (4.8%) answered that they traded water in 2009. As mentioned
Even though the number of observations on water transactions
above, none of the households interviewed in the baseline sur-
is rather limited, and is insufficient to do any meaningful statis-
vey covering the year 2007 answered that they traded water. The
tical analyses, a closer look at the available information on these
15 farm households engaged in water transactions in our study
7 exchanges may provide some useful insights. Table 3 provides more
are located in nine different villages in six different townships.
detailed information about the water transactions undertaken by
Among these 15 households, seven are living in one irrigation area,
each of the 15 households that were involved in water transfers in
the Hongshuihe irrigation area. With 105 of the 315 interviewed
2009. As can be seen from the table, 12 out of the 15 households
households living in the Hongshuihe area, this means that 6.7% of
that transferred water were actually swapping water without pay-
the interviewed households in that area answered that they traded
ments. That is to say, a household received an amount of water
water in 2009.
from another household in one round of irrigation and returned the
These data may be compared with those collected in an ear-
same amount of water in another round, or even within the same
lier study on tradable water rights in Zhangye City reported in
round. Such water swaps normally occur between households that
Zhang et al. (2009). In that study, survey data covering the year
are very familiar with each other. As discussed in Section 2, such
2003 were collected in five different irrigation areas (out of the 25
(non-market) reciprocal water use arrangements are very similar to
main irrigation areas in Zhangye City). Only one of those irrigation
labour-sharing arrangements that are frequently observed in peas-
areas, Hongshuihe, is located in Minle County; the other four
ant agriculture. The quantities of water that were exchanged in this
are located in other counties. Zhang et al. (2009) find no water
3 3
way varied from 10 to 540 m , with an average of 280 m .
exchanges in the four irrigation areas where there are groundwa-
Trading through water markets is almost non-existent. Two of
ter sources, and a very limited number of water exchanges in the
the 15 households bought water from another household (at a price
Hongshuihe irrigation area, where use of groundwater is not a real-
3 3
of 0.1 RMB/m , almost equal to the fee of 0.094 RMB/m paid for
istic alternative due to the high pumping costs involved. Among the
the allocated water) in one irrigation round in 2009, and sold the
380 households in the Hongshuihe irrigation area interviewed in
3
same quantity of water (100 and 150 m , respectively) to the same
their survey, five households traded water with other households.
household at the same price in another irrigation round in 2009. So
All transactions took place within the own village, and against
3 3 in fact they were also swapping water, but against a fixed payment.
payment. The average price was 0.025 USD/m (or 0.20 RMB/m ),
3 Only one household that swapped water with another household
while the average traded quantity amounted to 123 m .
3
also bought a large quantity of water (3000 m ) later in the season
Table 2 summarizes the trend in water transactions in the Hong-
3
at a price of 0.2 RMB/m . And finally there was one household that
shuihe irrigation area based on Zhang et al. (2009) and our two
3
received 900 m of water without payment in 2009 and also bought
surveys. The data suggest a slightly increasing trend over time, but
3 3
some water (100 m ) at a price of 0.16 RMB/m early in the year.
7
The nine villages (and the number of households involved) are: Yangjiajuan (2),
8
Wubacun (1), Erzhaicun (1), Yushumiao (2), Mazhuangcun (2), Wujiazhuang (2), Out of the 315 households in our sample, 135 households (i.e. 43%) were involved
Sanpucun (1), Wulangcun (3) and Zhujiangzhuang (2). in land transfers in the year 2009.
L. Zhang et al. / Agricultural Water Management 131 (2014) 70–78 75
Table 3
Traded water quantities and prices in Minle County, 2009.
3 3 3 3
Household Atlantic potatoes Direction Quantity (m ) Price (RMB/m ) Time Direction Quantity (m ) Price (RMB/m ) Time
grower
1 no in 150 0 NA Out 150 0 NA
2 yes in 440 0 12 June Out 440 0 12 July
3 no in 10 0 April Out 10 0 June
4 yes in 100 0.1 July Out 100 0.1 August
5 no out 350 0 August In 350 0 September
6 yes in 240 0 12 June Out 240 0 17 June
in 3000 0.2 20 Sept.
7 yes in 400 0 Early July Out 400 0 End July
8 yes out 150 0.1 May In 150 0.1 June
9 yes in 250 0 May Out 250 0 May
10 no in 25 0 3 June Out 25 0 10 June
11 yes in 500 0 May Out 500 0 May
12 yes in 350 0 August Out 350 0 2010
13 yes in 540 0 20 June Out 540 0 20 Sept.
14 no in 900 0 June In 100 1.6 January
15 yes in 100 0 5 June Out 100 0 20 June
NA = not available.
This raises the question why almost all of the (rather limited Atlantic potatoes growers were relatively more involved in water
number of) water transactions that we observed took the form of transactions than other farmers.
reciprocal water sharing arrangements. By exchanging water, both As mentioned before, Atlantic potatoes require a relatively large
parties can benefit from seasonal variations in water values. But amount of water, but need it later in the season. They do not
that would also be the case if water transactions take place through need water during the first irrigation round (early May), when
a market. As suggested by Tu and Bulte (2010), levels of general the seedlings are still small, but receive water during the other
and local trust may play a role in deciding to use either market irrigation rounds and extra irrigation rounds later in the year. As
or non-market exchanges. Although farmers in our research area discussed in Section 2, water users have an incentive to trade water
possess tradable water use rights, they do not possess certificates whenever the marginal net benefits of water differ between water
of those rights. Trading of water may therefore easily cause conflicts users. For Atlantic potatoes growers, the marginal benefits of water
between buyers and sellers. The perceived risks of water transac- are expected to be relatively low during the first irrigation round
tions depend on the extent to which both sides believe that the and much higher during the later rounds. In our data, however,
other side can be trusted. we hardly find any Atlantic potatoes growers who provide water
The baseline survey carried out in May 2008 contained ques- to other households during the first round (early May) and obtain
tions about trust. Respondents were asked to indicate their trust water in return during a later irrigation round (see Table 3). The
level to 10 different groups (parents, brothers/sisters, children, spring period is a very dry season in Minle County, and hence
other relatives, local officials, classmates/peers, neighbours, famil- there is a high demand for applying water to most crops during
iar (known) people, staff in companies, strangers), using a scale the first irrigation round. It may therefore be assumed that the
from 0 (totally distrust) to 1 (fully trust). Ma (2013) performed fac- marginal benefit of water is very high for all farmers during the
tor analysis, using the standard method of principal-component first irrigation round, and no water transactions take place. Later in
factors, to analyse the answers to these questions. Based on rotated the year, when there is more precipitation, the marginal benefits
factor loadings (pattern matrix), he distinguished between kin- of irrigation water become lower. The water that Atlantic potatoes
ship trust (trust to parents, brothers/sisters and children), trust farmers receive during extra irrigation rounds may not always fit
towards known people (the trust towards classmates/peers, neigh- the growing requirements of the crop. Hence, it makes sense for
bours and familiar (known) people) and trust towards strangers them to swap part of this additional water with other households
(trust towards staff in companies and strangers). The observed in their own WUG or WUA.
average levels of trust equal 0.96 for kinship trust, 0.79 for trust Why did only ten of the Atlantic potatoes growers exchange
towards known people, and only 0.44 for trust towards strangers. water in 2009, and the other 95 did not? A number of factors may
Ma (2013) uses these results to explain the segmented and under- explain the low incidence of water transactions. Firstly, knowl-
9
developed land rental market in the region. Our finding that the edge may be an important factor. Households with better-educated
limited number of water transaction in the region are reciprocal heads may be more efficient farmers that obtain higher marginal
transactions between households that are very familiar with each net benefits of water. The average education level of the head of
other, while impersonal market transactions are virtual absent, the household for all the households who planted Atlantic pota-
may be explained in a similar vain. toes in 2009 and did not exchange water is 7.95 years, while that
What role did the establishment of the potato processing com- of Atlantic potatoes growers who exchanged water in 2009 is 8.50
pany and the priority given to Atlantic potatoes in water allocation years on average (see Table 4). But the difference is not statistically
play in promoting these water exchanges? Among the 15 house- significant.
holds who were involved in water transactions in 2009, ten Secondly, a larger area planted with Atlantic potatoes results
households (i.e. 2/3) grew Atlantic potatoes that were supplied in the allocation of more water due to the water allocation pri-
to the new company. Out of the 315 households that we inter- ority policy, and may induce water selling. It may also induce
viewed, 105 (i.e. only 1/3) grew Atlantic potatoes in 2009. Hence, water swapping if the additional water is delivered at times when
the marginal benefits it provides to Atlantic potatoes farmers are
lower than those in other irrigation rounds and are lower than the
marginal benefits it offers to other farmers during the same round of
9
The regression results presented in Ma (2013) show that trust towards known
irrigation. In 2009, the size of the area planted with Atlantic pota-
people has a significant positive effect on the probability and intensity of land renting
toes for the ten Atlantic potatoes growers that exchanged water
and on the use of informal rental contracts.
76 L. Zhang et al. / Agricultural Water Management 131 (2014) 70–78
Table 4
Characteristics of Atlantic potatoes growers.
Education of head (years) Arable land (mu) Atlantic potatoes area (mu) WUR land (mu)
Mean Std. Dev. Mean Std. Dev. Mean Std. Dev. Mean Std. Dev.
Atlantic potatoes growers 8.00 2.42 20.0 11.2 0.92 0.657 15.3 11.7
**
Exchanging water 8.50 2.97 22.0 12.4 1.04 0.847 20.0 7.0
**
Others 7.95 2.91 19.8 11.1 0.91 0.638 14.8 7.6
Whole sample 7.49 3.51 19.6 12.5 0.31 0.575 13.8 7.8
**
Indicates that the difference between the group means is statistically significant at the 5% testing level. The differences between the mean values of the other variables
are not significant at the 10% level for farmers exchanging water and other Atlantic potatoes growers.
was 1.04 mu while it was 0.91 mu for the other farm households same WUA tend to grow similar crops with similar (planting and
growing Atlantic potatoes (Table 4). But again the difference is not irrigation) technologies.
statistically significant. Likewise, the difference in the total arable Costs of obtaining (reliable) information are an important ele-
land size did not differ significantly between the two groups in ment of transaction costs. In our survey held in May 2010 we
2009. One important aspect that differed significantly (at a 5% test- asked questions about farmers’ understanding of the tradability
ing level) was the size of the WUR land (see the last two columns of of their WUR in 2007 and 2009. Although the exchange of WUR
Table 4). Atlantic potatoes farmers who exchanged water had sig- is permitted since 2002 in this area, only 9.8% of the interviewed
nificantly more WUR land (22.0 mu on average) than other Atlantic households were aware that they could buy or sell water against
potatoes farmers (14.8 mu on average). As mentioned in Section 3, payment, while 24.1% realized that they were allowed to swap
water is allocated to crops other than Atlantic potatoes according water use rights in 2007. In 2009, these percentages were only
to their WUR area. Therefore, households owning land with a large slightly higher: 10.8% thought that they were allowed to buy or
WUR area receive more water for irrigating all their crops than sell water, and 27.9% mentioned that they were allowed to swap
10
households with relatively small WUR areas, and may find it prof- water (Table 5). In conclusion, these data suggest that informa-
itable to swap part of this water with other households to make it tion asymmetry between the government and water users may be
fit better to their crop growing requirements. one of the main obstacles to water trading in this region.
An important aspect of the functioning of water markets is the As mentioned above, the price that farmers can charge for water
3
pricing of water. Under a proper water trading scheme, the water should not exceed 0.3 RMB/m . Theoretically, the price constraint is
price should reflect its marginal value to buyers and sellers. In the expected to affect water market transactions by limiting the sup-
absence of transaction costs, each unit of water will be worth the ply of water and increasing its demand. However, in our survey
same at the margin to each agent after the exchange. For this to we found that only six respondents were aware of the upper limit
happen, water users should be free to set their own water prices. set by the government. Therefore, the price limit does not seem to
However, water prices are not fully market determined in Minle constrain market exchanges of WUR in this case.
County. According to the Bureau of Water Resource Management Given the current low water price and limited availability of
in Zhangye City, the prices of exchanged water are not allowed to water, the marginal benefits of water may exceed the actual water
3 3
exceed 0.3 RMB/m . If a household charges a higher price for its price (0.09 RMB/m on average) for many farmers. Zhang (2013)
WUR, the WUR allocated to that household can be withdrawn by estimated crop-specific marginal water values, based on a system
the local government. of crop production functions, for the 315 farm households that
High transaction costs may also be an important obstacle to the we interviewed in Minle County. The estimated marginal values
3 3
development of water markets (see Section 2). Transaction costs equal 0.046 RMB/m for grains, and 0.266, 0.543 and 0.567 RMB/m
faced by households interested in trading water include time and for Atlantic potatoes, other potatoes and other cash crops, respec-
costs involved in acquiring information on possible water trad- tively. In other words, crop-specific marginal values are far from
ing procedures, in searching for households willing to sell or buy being equal and there remains much scope for gains to be obtained
water, in negotiating the conditions of the water transfer, and in from water trade, especially between grain-oriented and cash crop
monitoring and enforcing water transfers (see also Zhang et al., oriented farmers.
2009). As discussed above, trust may play an important role in We can conclude that the low water price, the lack of appro-
this respect. Low levels of trust result in relatively high transac- priate information and the low level of trust towards unknown
tion costs. The water exchanges that we observed in our survey persons seem to limit the functioning of the water market, even
all occurred between relatives or neighbours, where levels of trust though a tradable WUR system has been introduced through the
tend to be high. pilot project and despite the fact that the establishment of the large
Information that we obtained during field visits and informal potato processing company has created more economic incentives
talks with farmers provides additional insights into the reasons for for water trading. If farmers would be well-informed that they are
the almost complete absence of market transactions. In the first allowed to exchange WUR, if they would be free in choosing a water
place, the amount of water allocated to households is often con- price without any restriction, and if appropriate conflict resolution
sidered insufficient for irrigating all the crops that they planted, mechanisms would be in place, more farmers will be expected to
let alone that they would have redundant water for selling. This sell or buy water at prices that most likely will exceed the current
implies that it is difficult for households willing to buy water to fixed water price and in some cases perhaps even the current ceiling
find potential sellers, given the prevailing water price ceiling. In on the water price.
the second place, if there exist large differences in marginal bene-
fits of water between farm households, the incentives for trading
water will be large. In our research area, however, differences
10
Most households that answered positively to these questions for the year 2007
in marginal benefits between farm households may be relatively
also gave a positive response for the year 2009: 9.2% (buying or selling) and 21.3%
small. Irrigation requirements mean that farmers need to tune
(swapping) respectively. In addition, the majority of the households that answered
their crop choice and management decisions with other households
positively to the question about buying or selling also answered positively to the
within the same WUGs and WUAs. As a result, farmers within the question about swapping water: 8.3% for the year 2007 and 8.6% for the year 2009.
L. Zhang et al. / Agricultural Water Management 131 (2014) 70–78 77
Table 5
Farmers’ understanding of water exchange rights.
Are you allowed to swap water with others?
2007 2009
Yes (%) No (%) No idea (%) Yes (%) No (%) No idea (%)
24.1 52.4 23.5 27.9 48.9 23.2
Are you allowed to buy or sell water?
Yes (%) No (%) No idea (%) Yes (%) No (%) No idea (%)
9.8 64.4 25.7 10.8 62.5 26.7
6. Conclusion rental market development in rural China, concludes that improve-
ments in the rural legislative system and its implementation may
In this study we examine irrigation water exchanges in a region reduce the costs borne by farm households in protecting their
in China where households possess tradable water use rights, land rights via formal legal means, such as official meditation
where a major agro-processing company has recently been estab- and arbitration and going to the court (Ma, 2013, Chapter 5). The
lished, and where the local government intervenes in the allocation development of water markets based on impersonal transactions
of water to stimulate farmers to grow a cash crop for that com- is likewise expected to gain from a better protection of water use
pany. Despite these favourable enabling and driving factors, we rights through improvements in the legal system and its imple-
find that water market transactions are virtually absent. Instead, mentation. With respect to credit constraints, the current literature
we observe that reciprocal water use arrangements (water swaps) suggests that they can be an important driving force of reciprocal
have emerged at a limited scale. labour-sharing arrangements. Given that most water swaps that we
The case crop stimulated by the new company and the local gov- observed in our household survey in Minle County took place with-
ernment, Atlantic potatoes, needs a relatively large amount of water out payments, credit constraints may also be an important driving
for growing, but it needs it later in the season than many other factor. Our survey, however, did not include questions that would
crops grown in the region. The water swaps that we observed take allow us to examine this issue.
place between relatives or neighbours, with usually no payments Given the rather limited number of farmers involved in water
involved, and seems to be meant to improve the timing of water transactions in the research area of this study, further research in
applications to crops with different seasonal water requirements. this field should preferably use a dataset with sufficient obser-
A more detailed analysis of the observed water exchanges vations on water transactions to perform meaningful statistical
shows that a relatively large share of Atlantic potatoes farmers analyses in order to test the tentative conclusions that we draw
are involved, and that these farmers tend to have better access to about factors driving the development of the water market. Such
irrigation water than other farmers. We further argue that high research should preferably be carried out for multiple years, and
transaction costs, resulting from existing information asymmetry use appropriate methods to correct for the potential impact of
between the government and farmers and from low levels of trust in time-dependent factors, like variations in rainfall and water avail-
unknown people, severely limit the trading of water. In 2009 only ability and changes in the (global, national and regional) economic
27.9% of the interviewed households was aware that they were environment, which we could not do in our study.
allowed to swap water with others, while only 10.8% knew that Creating proper institutions for the development of water mar-
they were allowed to buy or sell water against payment. Out of kets, reducing transaction costs involved in using water markets
these, only a minority (1.9% of the entire sample) knew the maxi- (particularly through provision of more adequate information) and
mum price they are allowed to charge in trading water. The few paid removing existing restrictions on water prices are important pre-
transaction that we observed in our survey took place at prices that conditions for improving economic efficiency of water use. In the
were close to the price set by the government for standard water implementation of these measures, due attention should be given
allocations to farmers. High levels of trust towards kinship and to potential negative effects on the achievement of other important
known people as compared to unknown people seem to play a role policy goals such as reduced income inequality and maintenance
in our finding that the few water transactions that we observe in the of grain self-sufficiency. If such negative effects are found to occur,
region consist almost entirely of water swaps, a non-market insti- appropriate counteracting measures may need to be undertaken
tution similar to reciprocal labour sharing arrangements, instead of without compromising the goal of achieving more efficient water
market transactions. Hence, we can conclude that factors other than use in irrigated agriculture.
an improvement in the output market (such as producer ignorance,
centrally set prices, trust) need to be considered, if improvement
in the market for irrigation water is to occur. References
More research is needed into the root causes of these problems,
particularly the information asymmetry, in order to come up with Bromley, D.W., 1989. Institutional change and economic efficiency”. Journal of Eco-
nomic Issues 23, 735–759.
policy recommendations on ways to deal with this problems. An
Chen, Y., Zhang, D., Sun, Y., Liu, X., Wan, N., Savenije, H.H.G., 2005. Water demand
important potential explanatory factor that needs to be taken into
management: a case study of the Heihe River Basin in China. Physics and Chem-
account in such follow-up research is the extent to which local gov- istry of the Earth 30, 408–419.
Easterly, W., 2005. The White Man’s Burden. Penguin, New York.
ernment self-interests, particularly the workload that promotion
Ellis, F., 1993. Peasant Economics, second ed. Cambridge University Press,
of water trading brings to bureaucrats, plays a role in the cur-
Cambridge, UK.
rently existing information asymmetry between the government Gilligan, D.O., 2004. The Economics of Agricultural Labour Exchange with Evidence
from Indonesia. In: Ph.D. Dissertation. University of Maryland, College Park, MD, and local farmers.